This thesis present research work conducted in the field of retina image analysis. More specifically, the work is directed at the application of texture analysis technique for the segmentation of common retinal landmark and for retina image classification. The main challenge in this research is in identifying the suitable texture measurement for retina images. In this research we proposed the used of texture measurement based on Binary Robust Independent Elementary Features (BRIEF). BRIEF measure texture by performing an intensity comparison in a local image patch, thus it is very fast to compute and tolerant to any monotonic increase or decrease of image intensities, which makes the descriptor invariant to illumination. The performance of BRIEF as texture measurement is first shown in an experiment involving texture classification and segmentation using common texture datasets. The result demonstrates good performance from BRIEF in this experiment. BRIEF is next used in two applications of retinal image analysis, namely optic disc segmentation and glaucoma classification. In the former, we proposed the used of pixel classification using BRIEF as textural features and circular template matching to segment the optic disc. In addition, an extension of BRIEF called Rotation Invariant BRIEF (OBRIEF) is later proposed to improve the segmentation result. For glaucoma classification, we described two approaches for glaucoma classification using BRIEF/OBRIEF features. The first is based on determination of cup to disc ratio (CDR) and the second is classification using image features i.e. BRIEF features. Overall, our preliminary results on using BRIEF as texture measurement for retinal image analysis are encouraging and demonstrate that it has the potential to be used in retina image analysis.

Diabetic Retinopathy (DR) is one of the principal sources of blindness due to diabetes mellitus. It can be identified by lesions of the retina, namely microaneurysms, hemorrhages, and exudates. DR can be effectively prevented or delayed if discovered early enough and well-managed. Prior studies on diabetic retinopathy typically extract features manually but are time-consuming and not accurate. In this research, we propose a research framework using advanced retina image processing, deep learning, and a boosting algorithm for high-performance DR grading. First, we preprocess the retina image datasets to highlight signs of DR, then follow by a convolutional neural network to extract features of retina images, and finally apply a boosting tree algorithm to make a prediction based on extracted features. Experimental results show that our pipeline has excellent performance when grading diabetic retinopathy images, as evidenced by scores for both the Kaggle dataset and the IDRiD dataset. / Master of Science / Diabetes is a disease in which insulin can not work very well, that leads to long-term high blood sugar level. Diabetic Retinopathy (DR), a result of diabetes mellitus, is one of the leading causes of blindness. It can be identified by lesions on the surface of the retina. DR can be effectively prevented or delayed if discovered early enough and well-managed. Prior image processing studies of diabetic retinopathy typically detect features manually, like retinal lesions, but are time-consuming and not accurate. In this research, we propose a framework using advanced retina image processing, deep learning, and a boosting decision tree algorithm for high-performance DR grading. Deep learning is a method that can be used to extract features of the image. A boosting decision tree is a method widely used in classification tasks. We preprocess the retina image datasets to highlight signs of DR, followed by deep learning to extract features of retina images. Then, we apply a boosting decision tree algorithm to make a prediction based on extracted features. The results of experiments show that our pipeline has excellent performance when grading the diabetic retinopathy score for both Kaggle and IDRiD datasets.

Diplomová práca sa zaoberá detekciou a hodnotením skreslených snímok v retinálnych obrazových dátach. Teoretická časť obsahuje stručné zhrnutie anatómie oka a metód hodnotenia kvality obrazov všeobecne, ako aj konkrétne hodnotenie retinálnych obrazov. Praktická časť bola vypracovaná v programovacom jazyku Python. Obsahuje predspracovanie dostupných retinálnych obrazov za účelom vytvorenia vhodného datasetu. Ďalej je navrhnutá metóda hodnotenia troch typov šumu v skreslených retinálnych obrazoch, presnejšie pomocou Inception-ResNet-v2 modelu. Táto metóda nebola prijateľná a navrhnutá bola teda iná metóda pozostávajúca z dvoch krokov - klasifikácie typu šumu a následného hodnotenia úrovne daného šumu. Pre klasifikáciu typu šumu bolo využité filtrované Fourierove spektrum a na hodnotenie obrazu boli využité príznaky extrahované pomocou ResNet50, ktoré vstupovali do regresného modelu. Táto metóda bola ďalej rozšírená ešte o krok detekcie zašumených snímok v retinálnych sekvenciách.

This master thesis deals with detection of blood-vessel bifurcations in retinal images and its properties. There are explained procedure of taking photographs of retina by fundus camera, optical coherence tomography (OCT) and scanning laser opthalmoscope (SLO) and properties of fundus images are described. In this thesis are mentioned some effective thresholding methods and there are explained the most important morphological operations with binary images, as well as with grayscale images. Detected bifurcations are used for image registration with second-order polynomial transformation using corresponding bifurcations.